An Analysis of Salmonella Contamination in Broiler Chickens

Executive Summary

Click here to get the PDF and
Click here to get the charts

The bacteria Salmonella is the leading cause of food-borne illness in the United States¹ with nearly a million cases of salmonellosis attributed annually to meat and poultry consumption.² Of these, more than 14,000 of the victims are hospitalized and more than 400 die.³ The estimated total annual cost of all cases, foodborne and otherwise, of salmonellosis is about $2.46 billion (in 2006 dollars).⁴ Concern about the potential for pathogens, including Salmonella, to become resistant to antibiotics also is increasing.  Infections caused by antibiotic–resistant bacteria are more frequently associated with illness and death than those caused by bacteria that are not resistant.⁵

With this update to our 2006 report, Foul Fowl: An Analysis of Salmonella Contamination in Broiler Chickens, we show the continuing failure of the U.S. Department of Agriculture to protect consumers from Salmonella contamination of broiler chickens.

For both the original report and this update, Food & Water Watch used the Freedom of Information Act to obtain USDA’s Salmonella testing results. The original report used data from 1998 through 2005 and this update uses data for 2006 through January 2008. 
 
We are releasing this information for several reasons. First, citizens have a right to information that indicates how effectively their government is ensuring the safety of products that carry USDA’s seal of inspection. Second, consumers have a right to public information concerning the relative performance of poultry–producing plants under government inspection. Third, publication of the names of plants that have failed to meet the regulatory standard may create additional incentive for plants to improve the safety of their processes.

Background of "More Foul Fowl", March 2008.

Background

USDA’s Food Safety and Inspection Service (FSIS), the government agency responsible for inspecting meat and poultry, has established acceptable levels of Salmonella contamination for different types of meat and poultry.

In 1998, FSIS began enforcing a new Salmonella standard by testing raw and ground products for the presence of the pathogen. The purpose of the program was to use microbial sampling to determine when plants were not controlling food safety hazards in their production processes. Testing also was supposed to objectively indicate when industry and government were not fulfilling their food safety responsibilities.

The acceptable level of contamination established for each product was based on the average level of Salmonella present for each class of product during initial baseline tests. Consequently, FSIS acceptable percentage of Salmonella contamination varies greatly by product.

To determine the contamination level for a particular plant, FSIS tests a sample of the finished product each day the plant is operating until the requisite number of samples is taken. The number of required daily samples differs by species. So, for example, the testing period for a plant that slaughters steers and heifers would be as long as is necessary to collect 82 daily samples, whereas the testing period for a broiler chicken plant would be long enough to collect 51 daily samples. The length of the testing period is also affected by the frequency with which a plant produces the product.  Some small plants do not produce every day, and therefore the testing period is longer at these plants to collect the same number of samples. The contamination rate of a plant is the percentage of the daily samples that are contaminated with Salmonella >during the testing period. 

As a matter of practice, if the plant’s contamination rate does not exceed the regulatory standard, it will not typically undergo another testing period for approximately one year.  However, on January 28, 2008, FSIS announced that plants with half or fewer of the allowable number of positive samples in their last two testing periods would not be scheduled for more tests for 12 to 24 months.⁶ We believe this reliance on previous performance is misguided because circumstances in a plant can change very quickly. If the agency implements this proposal, consumers could be at greater risk because past data indicates that a plant with fewer than the allowable number of positive Salmonella test results in one period can fail to meet the performance standard in a subsequent testing period.  In other words, the data shows that good performance does not always last for the length of time that this new policy would allow between testing periods.  See plants marked with an asterisk in Chart 2 for examples of when this new policy could have delayed testing during periods of poor performance.    

Initially, FSIS established an enforcement program for its Salmonella program with actions getting progressively more serious with each additional failed testing period and culminating in withdrawal of inspection if the plant failed to comply with the Salmonella regulation for three consecutive testing periods. In 2001, however, a federal District Court ruled that the agency could not withdraw inspection solely based on a plant’s failure to meet the requirements of the Salmonella regulation. Since then, the agency reportedly increases the level of scrutiny at a plant with each successive Salmonella failure and may take enforcement action after considering the results of those investigations.

Overall, the results of the FSIS program have been disappointing. While the agency claims that the 1998 implementation of new “[inspection] and pathogen reduction programs represents one of the most significant changes in the regulation of the meat and poultry industry since the inspection program began in the early 1900’s,” ⁷ the percentage of broiler chickens found by the agency’s testing program to be contaminated with Salmonella actually increased from 10.83 percent in 1998 to 11.41 percent in 2006.⁸

Salmonella Testing in Broiler Chickens

To determine the contamination level for a particular broiler chicken plant, the agency collects daily samples until it has 51. For high volume plants that are producing five to seven days each week, this testing period usually takes between two and three months.

To collect a single sample, an FSIS inspector takes a carcass from the end of the plant’s production line, puts it in a sterile plastic bag with a chemical solution and agitates it for one minute. This solution is then sent to an FSIS lab to determine if the carcass was contaminated with Salmonella.

Since 1998, the agency has published only generalized data from the program, usually on an annual basis. In 2006, the agency announced it would begin publishing quarterly data and it issued the first quarterly report on June 23, 2006.   The agency also announced that, to increase incentives for plants to produce safe food, it would begin categorizing plants based on their success in meeting the regulatory standard and would begin publishing the status of individual plants on its web site. While FSIS repeated the same intention in a January 2008 Federal Register notice,  it has thus far failed to report individual plant status. 

Under the agency’s categorization scheme, Category 3 plants are those that fail, meaning that for broilers more than 12 of the 51 samples (24 percent) test positive; Category 2 plants pass with seven to 12 of the 51 samples (13 to 24 percent) registering a positive result; and Category 1 plants are those that pass, with zero to six of the samples testing positive for Salmonella (a zero to 12 percent contamination rate) in both the current and the previous testing set.

Methods

Methods from "More Foul Fowl", March 2008.

The Salmonella testing results for 2006 and 2007 were acquired using the Freedom of Information Act. We have analyzed numerous FSIS testing databases over the past 10 years and routinely discovered errors or gaps in the data. All of our calculations are based on the data in the records we received.

We received the agency’s testing records from January 1, 2006 through January 31, 2008.  Some of these records included the establishment name as well as the establishment number, while others included just the establishment number.  For those records with just the number, we found the names of the companies operating these plants in the “Meat, Poultry and Egg Product Inspection Directory” on the agency’s website. The directory does not provide information for a few of the plants we found in the test result database. We requested the names of these plants from the agency, but had not received them as of publishing this report.

 -top-

Findings

Findings from "More Foul Fowl", March 2008.

Click here to get the charts

Chart 1 shows that 27 poultry slaughter plants (out of a total of 189 facilities) of all sizes failed FSIS Salmonella test sets in 2006 or 2007. This means that more than 12 of the 51 samples –– 24 percent –– collected in a testing period at each slaughterhouse registered positive for the pathogen. The chart also identifies the number of positive samples during the testing period at each plant.

Chart 2 reveals the compliance history of any broiler slaughter plant that failed in 2006 or 2007. It shows gaps of more than a year between FSIS testing for Salmonella in some plants. The chart also illustrates the fact that passing in one test period does not mean that contamination levels won’t increase beyond the performance standard in the next period. For example, a Perdue Farms facility (#19112P) in Kentucky failed Salmonella testing in February 2007 with 32 positive results for Salmonella, up from 9 positives (a passing score) in December 2005. Mountaire Farms of Delaware, Inc. (establishment #3P), passed a testing set in October 2002. However, FSIS did not complete another testing period there until June 2006, at which time the plant failed, with 19 of the 51 samples testing positive for Salmonella. Because FSIS was not testing during the three and a half year interim period, there is no way to determine when the plant began marketing chickens that posed a greater risk to consumers than government standards ostensibly allow.

Other large plants that were not tested yearly before a failing set include establishments 325P (Texas), 4653AP (Iowa), 13456P (Arkansas), and 13485P (Louisiana).

Such results undercut the position of FSIS that passing facilities should not be retested for 12 to 24 months. With no government oversight and enforcement, previously good plants may allow themselves to produce unsafe food over extended periods of time, which obviously threatens consumer health and safety.

Recommendations

Recommendations from "More Foul Fowl", March 2008.

In light of these findings, Food & Water Watch recommends that USDA:

• Seek legislation that makes performance standards enforceable under the meat and poultry inspection statutes.
• Publish on its website Salmonella testing results for each plant on a quarterly basis, including the number of samples taken at the plant and the number that tested positive for Salmonella.
• Abandon the proposal that plants with less than half the acceptable rate of Salmonella in their last two testing periods not be scheduled for another testing period for 12 to 24 months.